A scalable re-ranking method for content-based image retrieval

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Image Re-Ranking Acceleration on GPUs

Huge image collections are becoming available lately. In this scenario, the use of Content-Based Image Retrieval (CBIR) systems has emerged as a promising approach to support image searches. The objective of CBIR systems is to retrieve the most similar images in a collection, given a query image, by taking into account image visual properties such as texture, color, and shape. In these systems, the effectiveness of the retrieval process depends heavily on the accuracy of ranking approaches. Recently, re-ranking approaches have been proposed to improve the effectiveness of CBIR systems by taking into account the relationships among images. The re-ranking approaches consider the relationships among all images in a given dataset. These approaches typically demands a huge amount of computational power, which hampers its use in practical situations. On the other hand, these methods can be massively parallelized. In this paper, we propose to speedup the computation of the RL-Sim algorithm, a recently proposed image re-ranking approach, by using the computational power of Graphics Processing Units (GPU). GPUs are emerging as relatively inexpensive parallel processors that are becoming available on a wide range of computer systems. We address the image re-ranking performance challenges by proposing a parallel solution designed to fit the computational model of GPUs. We conducted an experimental evaluation considering different implementations and devices. Experimental results demonstrate that significant performance gains can be obtained. Our approach achieves speedups of 7x from serial implementation considering the overall algorithm and up to 36x on its core steps.

Extensive computation of albedo contrast between martian dust devil tracks and their neighboring regions

We have developed a method to compute the albedo contrast between dust devil tracks and their surrounding regions on Mars. It is mainly based on Mathematical Morphology operators and uses all the points of the edges of the tracks to compute the values of the albedo contrast. It permits the extraction of more accurate and complete information, when compared to traditional point sampling, not only providing better statistics but also permitting the analysis of local variations along the entirety of the tracks. This measure of contrast, based on relative quantities, is much more adequate to establish comparisons at regional scales and in multi-temporal basis using imagery acquired in rather different environmental and operational conditions. Also, the substantial increase in the details extracted may permit quantifying differential depositions of dust by computing local temporal fading of the tracks with consequences on a better estimation of the thickness of the top most layer of dust and the minimum value needed to create dust devils tracks. The developed tool is tested on 110 HiRISE images depicting regions in the Aeolis, Argyre, Eridania, Noachis and Hellas quadrangles. As a complementary evaluation, we also performed a temporal analysis of the albedo in a region of Russell crater, where high seasonal dust devil activity was already observed before, comprising the years 2007-2012. The mean albedo of the Russell crater is in this case indicative of dust devil tracks presence and, therefore, can be used to quantify dust devil activity. (C) 2014 Elsevier Inc. All rights reserved.